Thesis Quentin SENANT

Ledoux's (1998) model is an influential theory of the visual neurocognitive processing of fear in humans. It proposes that the amygdala is a processing hub for fear-related stimuli and that a subcortical pathway involving the superior colliculi (SC), the pulvinar (PUL) and the amygdala (AMY) transmits coarse information faster than the cortex would transmit fine visual features. This pathway is thought to receive magnocellular information from the retina, i.e. achromatic information composed of low spatial frequencies, sensitive to luminance contrasts and to movements. However, the elements that led to the creation of this model come from studies carried out on auditory fear conditioning in rodents and from the neuroanatomy of the macaque monkey. Applying such a model to humans in the absence of any certainty as to the pathways involved remains risky, and it is therefore necessary to assess its plausibility. The existence of patients suffering from affective blindness (i.e. capable of perceiving emotions in the absence of an explicit description) demonstrates that striated visual pathways are not necessary for correct categorisation of facial emotions, suggesting the existence of alternative visual pathways. However, the majority of the studies conducted in humans suffer from several biases.
Firstly, in patients diagnosed as having affective blindsight, the subcortical pathways identified in neuroimaging and electrophysiology studies are likely to be the result of brain plasticity phenomena and may therefore not be as significant as in non-brain-injured individuals. Secondly, the study methods used in non-brain-injured individuals are mainly correlational and therefore do not allow causal inferences to be made about the involvement of the SC-PUL-AMY pathway in threat processing. In other words, while the hypothesis that an activation of the pathway to the amygdala is correlated with the correct classification of expressions of fear is permitted, it remains impossible to affirm that this pathway is indeed at the basis of this processing or that it is necessary for it. In addition, results obtained by the Brisbane team in TN patient with affective blindsight could call into question the usefulness of Ledoux's (1998) subcortical pathway and its early activation during the processing of fear-related stimuli. Scalp EEG electrophysiological recordings of TN have shown that the right anterior frontal regions can activate before the amygdala, which is thought to activate very early in front of frightened faces. Furthermore, it has recently been proposed that contrast equalisation may influence studies designed to differentiate the influence of spatial frequencies on the classification of threat-related stimuli.
To this end, we propose, in the first part, to study the effect of contrast equalisation and of spatial frequencies on the categorisation of expressions of fear vs. expressions of joy using EEG and fMRI data. Secondly, we propose to conduct a series of experiments using ecological emotional scenes instead of emotional facial expressions. Indeed, it is likely that arousal is a stronger determinant of the response of the amygdala than the emotional valence of the stimuli itself. Finally, we plan to reproduce these studies using artificial neural networks in order to study the diagnostic visual informations that are sufficient for these tasks.

Supervisors :
Martial MERMILLOD martial.mermilloduniv-grenoble-alpes.fr (martial[dot]mermillod[at]univ-grenoble-alpes[dot]fr)
Alan Pegna a.pegnauq.edu.au (a[dot]pegna[at]uq[dot]edu[dot]au) (Codirection)
Nathalie GUYADER nathalie.guyadergipsa-lab.grenoble-inp.fr (nathalie[dot]guyader[at]gipsa-lab[dot]grenoble-inp[dot]fr) (Co-encadrant) et
Frédéric DUTHEIL frederic.dutheiluca.fr (frederic[dot]dutheil[at]uca[dot]fr)(Co-encadrant)

Keywords : A subcortical route to the amygdala